样本问题

The aim of this paper is to suppress negative effect of noise on complex firing patterns of neurons and extract deterministic dynamical relationship underlying irregular interspike intervals. We first give a brief introduction to ISI data to be analyzed. Then a simple nonlinear noise reduction method based on chaos theory is presented. The method is finally used to analyze ISI data both from a theoretical model and experiments.

围绕如何来消除神经元峰峰间期序列中随机噪声影响从而提取出决定不规则性的确定性动力学关系这个问题，本文首先简要介绍峰峰间期序列样本的制备，然后着重讨论一个简单可行的混沌时间序列降噪方法的原理和算法实现，最终将该方法运用到神经元放电活动数值模拟和实验记录到的峰峰间期时间序列样本分析中。

In order to deal with the problem of face recognition with one sample per person, a method called sub-block Principal Component Analysis based on partitions of the sample is presented in this paper.

针对单样本人脸识别问题，本文提出了一种基于单样本切割的子模块主成分分析方法。

Essay 2: This paper examines the effect of obesity on earnings. It analyzes the 2001 National Health Interview Survey of Taiwan using the ordered probit model with sample selection in which earnings are defined as an ordered variable. We also transform the earnings from ordered variable to continuous variable and use the Heckman's two-stage model to estimate the effect.

第二篇：本文主要在探讨肥胖对收入的影响，并使用国家卫生研究院，民国90年国民健康访问调查资料(National Health Interview Survey， NHIS)来进行分析，并采两阶段的ordered probit with sample selection模型来进行估计，其中第一阶段为样本的就业选择问题，而第二阶段为样本在有就业的情况下，肥胖对劳动者收入的影响，而此收入为区间资料；另外，本文还将收入的区间资料转换为连续资料，以Heckman's two-stage的方法估计，并比较两估计结果。

With introducing the Proportional to Population Size Sampling technique into resample for the problem of specimen reconstruction effectiveness and rationality, a neural network ensemble method based on the PPS technique was proposed, which could improve the accuracy of individual network and the differences among individual networks; and a new dynamic selection approach of individual network in a neural network ensemble was put forward.

针对样本重组的有效性和合理性问题，将PPS抽样技术引入样本重构，提出了基于PPS抽样的集成神经网络算法，以提高个体神经网络的准确性与差异度，并实现动态选择个体神经网络的神经网络集成新方法。

Concerning these problems, this paper proposed a new reduction strategy for large-scale training sample set according to the analysis on the structure of the training sample set based on the point set theory.

针对这些问题，通过在点集理论的基础上分析训练样本集的结构，提出了一种新的支持向量机大规模训练样本集缩减策略。

In chapter one,the writing research the questions in sample rotation when sample with simple random sampling according the process of the sample rotation.

第一章 这章内容属于文献综述性质的，主要是按样本轮换的实施过程，讲述前人对简单随机抽样下的样本轮换理论的有关问题。

In each period, without the knowledge of the distribution function of the time between failures and the repair time, the PA estimate of gradient of performance function is put into the stochastic approximation algorithm which recursively optimize the threshold value of the flow control model in a manufacturing system.

在分析过程中，对故障时间和工作时间的分布不做任何假设，利用自适应控制的思想，将整个样本路径依据一定的规则分成有限长的时间段，在每一段时间内用扰动分析估计目标函数的样本导数，然后用随机逼近算法渐进地估计流程控制问题中的最优阈值，从而实现了阈值的自适应控制。

In this paper,samples as a sensitive issue in two kinds of random survey answered model,from two angles of a large sample theory and a small sample theory respectively,the estimates obtained by the quantity of unbiasedness ,effectiveness,consistency of the strength of the areas of evaluation and assessment are to survey it is reasonable and effective.

以敏感问题调查中两种随机化回答模型为例，分别从小样本理论和大样本理论两个角度，对所得到的估计量进行无偏性、有效性、强弱相合性等方面的评估，从而评价此种统计调查方法的合理性和有效性，最后给出两种模型的实例分析。

In solving the problem of trouble-locating in the case of samples with great difference in their number for their different varieties, the training with C-SVM was undesirably under bias towards those varieties with fewer samples, so that the training accuracy was unsatisfactory.

在解决故障检测等分类问题时，若不同类别样本数目相差很大，C-SVM训练的分类错误总偏向于样本数较少的类别，因而影响了分类的精确性。

Cooperating the advantages of SVM solve classification problems about small sample set, non-linear and multi-dimension model identification, I use SVM to train the sample set of different frequency signal and look for a super plane to divide the data into two parts. This method gets rid of complex computation which is a flaw of traditional decoding method.This paper introduces the theory and algorithm of SVM at first. It proposes to use the SVM to decode FSK signal based on the classifiable character of SVM.

结合支持向量机在解决小样本、非线性及高维模式识别问题中表现出的许多特有优势，运用支持向量机对不同频率信号的数据样本集进行训练，通过寻找最优超平面来对不同频率信号的数据进行快速分类，克服传统频移键控解码算法中需要进行大量运算的缺点，从而实现频移键控的解码。

I can not make it blossom and suits me

我不能让树为我开花

When temperatures are above approximately 80 °C discolouration of the raceways or rolling elements is a frequent feature.

当温度高于 80 °C 左右时，滚道或滚动元件褪色是很常见的特征。